The Impact of the Congestion Charging Scheme on Air Quality in London

EPA Grant Number: R834677C155
Subproject: this is subproject number 155 , established and managed by the Center Director under grant R834677
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Health Effects Institute (2010 — 2015)
Center Director: Greenbaum, Daniel S.
Title: The Impact of the Congestion Charging Scheme on Air Quality in London
Investigators: Kelly, Frank
Institution: Health Effects Institute (HEI)
EPA Project Officer: Chung, Serena
Project Period: April 1, 2010 through March 31, 2015
RFA: Health Effects Institute (2010) RFA Text |  Recipients Lists
Research Category: Health Effects , Air


The study of the London Congestion Charging Scheme (CCS), conducted by Professor Frank Kelly and colleagues, was funded under HEI’s research program aimed at measuring the possible health impacts associated with actions taken to improve air quality. With this research program, HEI has sought to (1) fund studies to assess the health outcomes associated with regulatory and incentive-based actions to improve air quality at local or national levels, and (2) develop methods required for, and specifically suited to, conducting such research.

The CCS offered an unusual opportunity to investigate the potential impact on air quality of a discrete and well-defined intervention to reduce traffic congestion in the middle of a major city. The CCS was implemented in London in February 2003 with the primary aim of reducing traffic congestion by charging vehicles to enter the central part of London, defined as the congestion charging zone (CCZ). In an earlier study based on data from the first year of the scheme, members of the investigative team had reported early findings of modest reductions in the number of vehicles entering the zone and had projected declines of about 12% in emissions of both PM10 (particulate matter with an aerodynamic diameter of #10 μm) and nitrogen oxides (NOx) within the CCZ. Recognizing that these projected reductions, coupled with the small area represented by the CCZ within Greater London, could lead to relatively small changes in air quality, the HEI Health Research Committee recommended that the investigators first assess the actual changes in air quality and postpone their proposal to study health impacts until the air quality studies were completed. The investigators proposed a multifaceted approach to exploring the impact of the CCS on air quality, which involved a variety of modeling techniques, analysis of air monitoring data, and a newly developed assay for the oxidative potential of PM.


Kelly and his colleagues undertook a stepwise approach to evaluating the impact of the CCS on air quality. In the first part of their study, they updated emission estimates that had originally been developed for the London transportation agency, Transport for London. Using the King’s College London Emissions Toolkit (a set of statistical models and data), they developed detailed estimates of NOx, nitrogen dioxide (NO2), and PM10 emissions from vehicular and non-vehicular sources throughout the London area for the 4-year period encompassing 2 years before (pre-CCS) and 2 years after (post- CCS) the introduction of the scheme on February 17, 2003. Vehicular PM10 emissions were predicted from two primary sources — tire and brake wear and exhaust. These emission estimates were then input to a modeling system (the King’s College London Air Pollution Toolkit), which the investigators used to predict annual mean ambient concentrations of NOx, NO2, and PM10 throughout London for each year of the study. They explored how various assumptions about the mix of vehicles, speed, and congestion over the study period might affect the predicted spatial patterns of changes in air quality associated with the implementation of the CCS.

The results of the modeling exercise were also used to help select the fixed, continuous air monitoring sites from the London Air Quality Network (LAQN) with which to evaluate measured changes in air quality. The investigators created a CCS Study Database consisting of validated (or ratified) measurements of carbon monoxide (CO), nitric oxide (NO), NO2, NOx, and PM10 from monitors sited to record roadside or urban or suburban background air pollutants across London. The investigators calculated and compared geometric mean concentrations of these pollutants for the 2 years before (2001–2002) and 2 years after (2003–2004) the scheme was introduced. The changes over time at monitors within the zone were compared with changes during the same period at similar classes of monitoring sites in a control area more than 8 km from the center of the zone. In addition, the investigators also explored three other analytic techniques for characterizing and evaluating both projected and measured changes in pollutant concentrations over the period of the study: ethane as an indicator of pollutant dispersion due to regional atmospheric conditions; the cumulative sum statistical technique to identify step changes in air pollution data; and specialized graphical techniques to improve the siting of pollutant monitors by characterizing the dependence of pollutant concentrations at potential monitoring sites on local emission sources.

In the second part of the study, the investigators set out to explore whether implementation of the CCS led to detectable changes in either the composition of the PM10 mixture or in its oxidative potential — an indicator of toxicity. As part of this analysis, they sought to establish a more comprehensive baseline of monitoring data to use in future studies of the CCS by collecting data from additional monitoring sites located within and outside a proposed expansion of the CCZ known as the Western Extension.

For these purposes, the investigators created an archive of about 730 filters from tapered element oscillating microbalances, a type of PM10 monitor used at 16 sites within and surrounding the CCZ, including the Western Extension; the filter archive covered the 3 years before and 3 years after the CCS was introduced. After extracting the PM from the filters, they measured the oxidative potential of the extracts using an in vitro assay that measures the ability of the extracts to deplete antioxidants in a synthetic respiratory tract lining fluid. The investigators’ focus on oxidative potential, a measure of the capacity to generate oxidation reactions, arises from a leading theory about the causal role that oxidative stress may play in the health effects associated with exposure to air pollution. Their goal, in essence, was to use oxidative potential as an indicator of the potential toxicity of PM and to evaluate how it varied across London and in response to the introduction of the CCS.

To study the composition of PM, each filter extract was also analyzed using inductively coupled plasma mass spectrometry for a panel of metals that have been associated with traffic sources in studies by other investigators. Additional experiments were done to understand the relative contribution of different metal and non-metal components of PM to the oxidative potential measured in the assays.

Expected Results:

The modeling studies predicted small changes in both emissions and ambient concentrations of NOx, NO2, and PM10 across London that could be related to the implementation of the CCS, although the effects within the CCZ were projected to be more pronounced than elsewhere. They projected somewhat larger average reductions (about 20%) in NOx and PM10 emissions than the 12% reductions that had been predicted in the initial feasibility studies that preceded the CCS. However, the difference in these projections may partly be explained by the fact that the modeling in this study compared the 2 years before and 2 years after the introduction of the CCS, whereas the earlier estimates had been based on an analysis of only the first year of the scheme (2003). The investigators reported that unusual meteorologic conditions had led to periods of elevated pollution levels in London during that year.

Despite the somewhat larger projected reductions in emissions, the projected changes in concentrations of NOx, NO2, and PM10 related to the CCS were small. Within the CCZ, the investigators projected a net decline of 1.7 ppb in the annual average mean NOx concentration and a decline of 0.8 μg/m3 in PM10. The modeling also suggested that a major proportion of PM10 might be accounted for by regional background levels, but that contributions from tire and brake wear might also be important. NO2 was projected to increase slightly, by 0.3 ppb on average; the investigators attributed this increase to higher NO2 emissions associated with the introduction of particle traps on diesel buses as part of Transport for London’s improvements in the public transport system.

From their comparison of actual air pollutant measurements within the CCZ with those at control sites in Outer London, the investigators reported little evidence of CCS-related changes in pollutant levels at roadside monitoring sites, where their modeling had suggested the most pronounced effects would be seen. The effects of the CCS were more evident at urban background sites within the CCZ when compared with concentrations at sites in the control area: PM10 concentrations declined by 12% at the one background site in the CCZ where it was measured, and NO declined by between 10% and 25% at the three background sites where it was measured. However, levels of NO2 increased by between 2% and 20% at the three background sites compared with levels at the control sites; these increases were consistent with the predictions from the modeling studies and with the likely effects of the parallel intervention that introduced more filter-equipped diesel buses. The investigators concluded that the small net changes in NOx detected at both roadside and background monitoring sites — likely resulting from reductions in NO offset by increases in NO2 — did not provide strong evidence of an impact of the CCS.

In the study of the oxidative potential of PM10, the investigators were unable to identify a temporal, CCSrelated change during the 6-year period that encompassed the implementation of the scheme. However, the city-wide spatial analysis of oxidative potential revealed that PM10 sampled from roadside locations showed greater oxidative activity than PM10 sampled at urban background sites.

When they coupled these spatial analyses of oxidative potential with analyses of the metal content of PM10 from the same filters, the investigators concluded that their results provided suggestive evidence that PM10 derived from tire and brake wear (indicated by the presence of the metals arsenic, barium, copper, iron, manganese, nickel, and vanadium) might contribute to the oxidative potential of PM seen in filters from roadside monitoring sites. However, the investigators noted that correlations among the concentrations of PM10 attributed to exhaust and to tire and brake wear made it difficult to isolate how much these individual sources might contribute to the oxidative potential of PM10. Their other experimental findings suggested that the non-metal components of PM10 did not contribute substantially to oxidative potential in this assay, but the investigators could not rule out a role for all other non-metal components of ambient air pollution.

Overall, the investigators concluded that their primary and exploratory analyses collectively suggested that the introduction of the CCS in 2003 was associated with small temporal changes in air pollutant concentrations within the CCZ compared with those in control areas thought to be beyond the influence of the scheme. In addition, they observed that a number of limitations, including concurrent changes in transportation and emission control policies, unusual meteorologic conditions the year the scheme was introduced, and the influence of strong local sources on particular monitors, would preclude them from attributing these changes to the CCS alone. They also acknowledged that the area covered by the CCS — approximately 1.4% of Greater London — was likely too small to influence air pollutant levels substantially either within or outside the zone.

Relevant Websites: Exit

Main Center Abstract and Reports:

R834677    Health Effects Institute (2010 — 2015)

Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R834677C149 Development and Application of a Sensitive Method to Determine Concentrations of Acrolein and Other Carbonyls in Ambient Air
R834677C150 Mutagenicity of Stereochemical Configurations of 1,3-Butadiene Epoxy Metabolites in Human Cells
R834677C151 Biologic Effects of Inhaled Diesel Exhaust in Young and Old Mice: A Pilot Project
R834677C152 Evaluating Heterogeneity in Indoor and Outdoor Air Pollution Using Land-Use Regression and Constrained Factor Analysis
R834677C153 Improved Source Apportionment and Speciation of Low-Volume Particulate Matter Samples
R834677C155 The Impact of the Congestion Charging Scheme on Air Quality in London
R834677C156 Concentrations of Air Toxics in Motor Vehicle-Dominated Environments
R834677C158 Air Toxics Exposure from Vehicle Emissions at a U.S. Border Crossing: Buffalo Peace Bridge Study
R834677C159 Role of Neprilysin in Airway Inflammation Induced by Diesel Exhaust Emissions
R834677C160 Personal and Ambient Exposures to Air Toxics in Camden, New Jersey
R834677C162 Assessing the Impact of a Wood Stove Replacement Program on Air Quality and Children’s Health
R834677C163 The London Low Emission Zone Baseline Study
R834677C165 Effects of Controlled Exposure to Diesel Exhaust in Allergic Asthmatic Individuals
R834677C168 Evaluating the Effects of Title IV of the 1990 Clean Air Act Amendments on Air Quality
R834677C172 Potential Air Toxics Hot Spots in Truck Terminals and Cabs
R834677C173 Detection and Characterization of Nanoparticles from Motor Vehicles
R834677C174 Cardiorespiratory Biomarker Responses in Healthy Young Adults to Drastic Air Quality Changes Surrounding the 2008 Beijing Olympics